GATE Scientific meeting 2023

24 avr. 2023, 09:00 → 26 avr. 2023, 14:00 Europe/Paris

Aurélien Coussat, David Sarrut (CNRS), Lydia Maigne (University Clermont Auvergne), Wojciech Jan Krzemien (National Centre for Nuclear Research, Krakow)

Gate Scientific Meeting 2023

From 24 to 26 April 2023, Kraków, Poland

We invite the GATE community to participate to our scientific meeting. The meeting will cover scientific challenges in PET, SPECT, CT imaging as well as internal and external radiation therapy.

The meeting will be held in person in Kraków, Poland, on two different sites: the FAIS campus of the Jagiellonian University, and the Center for Theranostics. You can find more information about the venue on the related page.

To participate, please register before April the 13th and do not hesitate to propose a presentation (title+abstract) concerning your last advances with the GATE platform.

Best regards,

The OpenGATE collaboration

---

leaflet-v5.pdf

lundi 24 avril

09:00 → 10:15 GATE pre-meeting hackathon: Hackathon part 1

10:15 → 10:30 Coffee break

10:30 → 12:30 GATE pre-meeting hackathon: Hackathon part 2

12:30 → 14:00 Lunch

14:00 → 15:15 GATE pre-meeting hackathon: Hackathon part 3

15:15 → 15:30 Coffee break

15:30 → 17:00 GATE pre-meeting hackathon: Hackathon part 4

mardi 25 avril

08:40 → 10:20 GATE scientific meeting 2023: Session 1

08:40 Welcome

Orateurs: Prof. Paweł Moskal, Ewa Stępień

09:00 Introduction to Gate for newcomers - discussion

Orateur: Dr Lydia Maigne (University Clermont Auvergne)

GATE_Introduction_Maigne_25-04-23.pptx

09:40 Gate Simulations of a Novel Positron Emission Tomography based on Liquid Opaque Detection

The main experimental challenging developments in positron emission tomography (PET) are currently focusing on improving the sensitivity either by pushing the time of flight (ToF) performances or extending the axial field of view. As an alternative the use of 3γ detection capability enables several extra advantages, including the possible enhancement of the ToF information to better locate the annihilation vertex or enabling tissue probing through positonium study. We propose the LiquidO-PET (LPET) projet which aim to develop of a demonstrator to explore our ability to meet the aforementioned capabilities through a novel PET system paradigm. It is based on opaque scintillation of a liquid medium. The enormous light scattering forces the emission photons to stay around their source into a stochastically confined trajectory or “light ball” with a radius in the order of the centimeter and collected and mapped through a lattice of wavelength-shifting fibers embedded in the scintillator and read out by silicon-photomultipliers coupled to sub-100 ps readout electronics on both ends. The liquid scintillator medium with low-z and low density, causes the γ detection to be largely dominated by Compton scattering. Thus, a 511 keV γ manifests a sequence of several Compton scatters up to 50 cm from the injection point, each scatter leading to a point-like light ball, thus enabling the Compton interactions to be characterized by two specific signatures: i) the clear identification of the first Compton vertex (FCV), which is most precious for PET imaging and ii) the full Compton tracking (FCT) providing complementary tracking information for the full event as well as allowing 3γ detection capabilities. We present here the GATE simulations of the preliminary design and the projected performances of the demonstrator based notably on optical simulations in a diffusive medium. Expected performances range from a radial spatial resolution of a few mm to a few cm in the axial direction, a ToF resolution of 100-300 ps and a sensitivity of 200-700 kcps/MBq for an axial FoV in the 50-200 cm range.

Orateur: Dr Marc-Antoine Verdier (IJCLab - Université Paris Cité)

LPET_GATE_Krakow2023_MAV-1.pdf

10:00 General Electric Signa PET-MR simulations: validation with GATE

PET imaging is undergoing continuous improvements thanks to developments of new
detection approaches, new designs and new electronic chains. For example, whole-body
scanners' geometries, advances in detectors' time resolution, and improvements in TOF image reconstruction are hot topics for dynamic acquisitions.
Monte Carlo GATE simulations can be used to quantify the impact of these improvements
before or with prototype experimentations. The simulation must be validated against state-of-the-art imaging devices to produce realistic data. In this work, we model the GE SIGNA PET-MR system using GATE 9.1. The image reconstruction, including normalization and
attenuation corrections, is done with CaSTOR 3.1.1.

We present the simulation validated for spatial resolution, sensitivity, noise equivalent count rate (NECR) and image quality with and without TOF information. Furthermore, by using improved coincidence time resolution (CTR) promised by new detection prototype systems, we show the potentially achievable impact on image quality for clinical applications. We also discuss the perspectives of a whole-body scanner simulation based on the GE SIGNA PET-MR technologies and other new PET detector approaches.

Orateur: M. Adrien Paillet (BioMaps-Université Paris-Saclay-SHFJ-CEA)

OpenGate2023Biomaps_V2.pdf

10:20 → 10:50 Coffee break

10:50 → 12:10 GATE scientific meeting 2023: Session 2

10:50 Towards Total-Body J-PET

Orateur: Wojciech Krzemien (National Centre for Nuclear Research)

krzemien_gate_short.pdf

11:10 GATE simulated phantom studies on Positrigo's NeuroLF, Siemens' Biograph Vision and Quadra

Orateur: M. Philipp Mohr (University Medical Center Groningen)

GATE_meeting_2023_Groningen-1.pdf

11:30 Simulation of a neuroimaging experiment with MAPSSIC, a β+ sensitive intracerebral microprobe

The correlation of molecular neuroimaging and behavior studies in the preclinical field is of major interestto unlock progress in the understanding of brain processes and assess the validity of preclinical studies in drug development. However, fully achieving such ambition requires being able to perform molecular images of awake and freely moving animals whereas currently, most of the preclinical imaging procedures are performed on anesthetized animals. To achieve such a combination, we propose MAPSSIC, a pixelated intracerebral probe based on the CMOS technology implantable in rat brains. This probe provides real time images of β+ radioisotopes concentrations along its axis. Thanks to its in situ position, the probe is able to directly detect positrons. We present the simulations
with GATE of a typical animal experiment using a widely used β+ radiotracer with a fully voxelized rat phantom to assess the capacities expected with the probe.
We show that, at the optimal radiotracer distribution time point, 99.52 % ±0.10 % of the events recorded in the specific probe come from the skull area. Moreover, nearly 90 % of the events are from a direct detection of positrons which ensures the recording of local information.

Orateur: Samir El ketara (IJCLab, Université Paris-Saclay)

MAPSSIC_El_ketara_GATE_meeting_2023.pdf

12:10 → 14:00 Lunch

14:00 → 15:40 GATE scientific meeting 2023: Session 3

14:00 Using Gate to predict clonogenic alteration in radiosensitization studies

As part of the european project RXNanoBRAIN, the image-guided radiation therapy and the radiosensitization by Gadolinium and Bismuth hybrid nanoparticles are investigated for the treatment planning of the glioblastoma brain tumor. In this context, we performed Gate Monte Carlo simulations to assess the deposited dose in clusters of Bismuth-Gadolinium nanoparticles in U251 glioblastoma cells. Ionization process, created secondary particles and their maximal range were analyzed in the presence of nanoparticles. We also studied the mount of ionization into cells nucleus in presence or absence of nanoparticles. In parallel, we performed biological experiments, including
DNA damage assessment by immunofluorescence technique and clonogenic assay (from 0 up to 10 Gy). Then, biological and numerical results were compared and we observed that radiosensitization was not due to a DNA damage increase. We also propose a general prediction model taking into account the nanoparticle concentration and the planed deposited dose in water. Further biological alterations in these condition are currenlty investigated to better understand the radiosensitization process to include it into a more biological dosimetry model.

Orateur: Mlle Sarah Blind (CRAN, BioSiS, Université de Lorraine)

Slides_Sarah_Blind.pdf

14:20 BioDose actor implementation and validation in GATE9.3 and GATE10

Orateur: Dr Alexis Pereda (LPC UMR 6533 CNRS)

BioDoseActor_Pereda_25-04-23.pdf

14:40 The Cell Population platform (CPOP), Geant4DNA chemistry module

Orateur: Dr Lydia Maigne (University Clermont Auvergne)

CPOP_chemistry_Maigne_25-04-23.pdf

15:00 Monte Carlo Simulation of Elekta Geometry: Effect of Flattened Filter on Dose Distribution in a Heterogeneous Phantom

In this study, we aimed to investigate the effect of flattened filter on dose distribution in a heterogeneous phantom using Monte Carlo simulation. The GATE Monte Carlo code was used to simulate photon beams of 6 MV in a heterogeneous phantom consisting of lung and breast tissues. The electrons accelerated by 6MV in flattened and unflattened modes were configured in Gaussian parameters. The accuracy of the simulations was validated by comparing the results with experimental measurements. All Monte Carlo simulations were performed using the HPC-Marwan high computing grid, which allowed us to simulate
several parallel jobs. The statistical uncertainty was less than 2% for all results with high correspondence between the simulations and experimental measurements. The results showed that the dose in the flattened filter-free (FFF) mode was significantly higher compared to the flattened filter (FF) mode.

Orateur: M. Mustapha Assalmi (Universite Mohammed First)

15:20 The benefits of particle transport simulations in a web browsers.

Orateur: Dr Leszek Grzanka (IFJ PAN Krakow)

GATE_LG_Leszek_Grzanka.pdf

15:40 → 16:10 Coffee break

16:10 → 17:50 GATE scientific meeting 2023: Session 4

16:10 Recent developments of CASToR at J-PET: an overview for GATE users

Orateur: Aurélien Coussat (Jagiellonian University)

gate_meeting_2023_coussat.pdf

16:30 New Digitizer in Gate 9.3

Digitizer in the current GATE plays an important role for PET, SPECT, CT, Compton Camera detectors or more generally for imaging system modeling. It simulates the response of the photodetection components by a sequence of analytical models. Digitizer takes as input a list of interactions, Hits, within a crystal or detector element and generates digital pulses. In a digitization, the last one represents a detector output, such as an ADC/TDC count or a trigger signal. The first step of Digitizer is to construct Singles with a help of a set of dedicated modules, each of them mimics one specific hardware response with their associated uncertainty. In the case of PET imaging, a specific Digitizer part for constructing Coincidences exists.
Digitizer was written for the first versions of GATE, about 20 years ago. Since then it has grown in a part of the code that can be hardly maintained: some of the digitizer modules are not used anymore or duplicated by mistake and some of functionalities are not in working order anymore. Therefore, its update is required in order to adapt it to the novelties of Geant4, to clean up its current version and add new features.
In this talk, the GATE New Digitizer implementation (since version 9.3) will be presented. Added functionalities, the impact of changes on users, current status of the work and perspectives will be discussed.

Orateur: Olga KOCHEBINA

OKochebina GND GateMeeting Krakow 2023.pdf

16:50 Gate 9.3 and 10.beta releases

Orateur: David Sarrut (CNRS)

python gate status 2023.pptx

19:30 → 21:30 Dinner

mercredi 26 avril

09:00 → 10:20 GATE scientific meeting 2023: Session 5

09:00 Dose activities boosted with AI from lab LaTIM

Dose distribution is simulated by Monte Carlo (MC) method generally. However, due to its data dependency, complex model, long computational time, instead, our work utilize AI models to predict dose distribution in real time, which is a series of mathematical operations that fit on a wide range of data and flexible in handling new cases. Furthermore, this work compares the MC simulation (OpenGate based on Python) and DL methods (3D U-Net) for predicting dose maps. And the preliminary results are promising.

Orateur: Dr Jing Zhang (LaTIM, UBO)

OpenGate-presentationJing-LaTIM.pdf

09:20 Monte Carlo research activities at the Institute of Nuclear Physics PAS in Kraków

Orateur: Dr Jan Gajewski (Institute of Nuclear Physics PAN, Krakow, Poland)

Presentation Jan Gajewski.pdf

09:40 GATE RTion v1/IDEAL v1 for independent dose calculation in carbon ion beam therapy

The IDEAL project aims to substitute the patient specific quality assurance (PSQA) measurements for light ion beam therapy by independent dose calculation (IDC) using Gate-RTion. In this contribution, the commissioning status of the IDEAL v1 at MedAustron Ion Therapy center will be discussed including a physics settings variation study. Further, the status of the integration of IDEAL v1 into the commercial software myQA-iON (IBA Dosimetry, Germany) will be presented.

To find an optimal balance between calculation time and dose calculation accuracy, physics settings in GATE-RTion were varied and compared to dose commissioning data of the MedAustron carbon ion beam line. Four reference nuclear physics lists, two EM physics lists and secondary particle settings were compared against 16 measured depth dose profiles in water, lateral profiles at various distances in air and absolute dose in reference condition, respectively, with and without range shifter. Calculation times varied considerably for the different settings, while the dose calculation accuracy was in all tested cases within tolerances. Combining the fastest options could speed up the simulations by up to a factor 10 in some scenarios compared to the current default settings in IDEAL 1.

Orateur: Andreas Resch (MedAustron Ion Therapy Centre)

Resch_Gate_Meeting_2023_IDEAL1_v2_noanimations.pdf

10:00 Verification of proton treatment planning by means of Monte Carlo

In Proton Therapy Center in Prague we have developed a method for patient based quality assurance. Our project is based on Monte Carlo (MC) and is built on Gate. By means of MC, dose computational algorithms of local treatment planning systems XiO 5.1 (Elekta, Sweden) and RayStation 11B (RaySearch, Sweden) can be verified as MC method provides the highest calculation accuracy. It provides complementary information to QA measurement in water as dose distribution is calculated in real inhomogeneous patient environment. Linear energy transfer distribution in patient can be simulated alongside. High computation time is reduced through parallel computing with HTCondor (https://htcondor.org) – an open-source high-throughput computing software. Currently we compute on 204 CPU cores.

Orateur: Dr Klara Badaraoui Cuprova

Verification of proton treatment planning by means of MC.pptx

10:20 → 10:50 Coffee break

10:50 → 12:10 GATE scientific meeting 2023: Session 6

10:50 In-beam and post irradiation measurement of proton treatments with the PETITION PET detector

The PETITION PET detector, has been developed for in-beam use at Gantry 2 at PSI,
with an open ring design allowing for measurement of patient activation both during and
following proton therapy treatment. Tools for time resolved simulations of patient
treatments have been implemented using Gate. Intra-field imaging in the dead time
between delivery of each energy layer, approximately 100 ms, can be performed for
measurement of short-lived isotopes such as 12N (t1/2 = 11 ms). Following this inter-field
imaging of the produced isotopes, primarily 15O and 11C (t1/2 = 122 s and 20.3 min
respectively), can be performed between delivery of subsequent fields to generate 3D
images of the induced activity distribution. We also consider a scenario in which the PET
scanner is rotatable, moving the opening of the ring to allow for delivery of each new
treatment field. If coincidences are recorded while the scanner rotates, the loss of spatial
resolution due to the opening of the scanner may be compensated, allowing for possible
detection of anatomical changes in the patient. Imaging for approximately 30 seconds,
during the couch and gantry repositioning, provides sufficient counts for reconstruction of
the activity. The activity fall off of short lived isotopes produced during delivery of the first
treatment field is well correlated to the dose fall off (R2=0.947). Imaging
of the field while rotating the scanner leads to a reduction in the mean-absolute-error by
54-70% compared to imaging at a fixed angle for each field. Development of a new PET
reconstruction software for the purpose of inter-field imaging allows for fast reconstruction
times, and could potentially provide validation of treatments within clinically relevant time
frames. In beam measurements with phantoms are currently being performed, and
comparison with Gate simulations is ongoing.

Orateur: Keegan McNamara (Paul Scherrer Institute)

Keegan McNamara -Opengate scientific meeting.pptx

11:10 Towards GATE 10 for ion beam therapy applications

In ion beam therapy, Monte Carlo (MC) particle transport simulations are being increasingly used to perform independent dose calculation (IDC) for quality assurance in light ion beam therapy. In practice, IDC consists in recomputing the dose in the patient starting from the treatment plan specifications, with a Monte-Carlo engine. The results can be compared with the planned dose as a verification of the treatment before delivery to the patient.

As a carbon ion treatment plan (TP) consists typically in the order of several 104 single Pencil Beams (PBs), small errors in the description of a single PB may add up to a considerable discrepancy in the treatment plan. Therefore, first we implemented the PB source into GATE 10. A single PB is characterized with 3 parameters to describe the particle type, mean energy and energy spread and 8 parameters to describe the (correlated) beam optics (spot size, divergence, emittance, convergence flag; each for the x and y direction). The implementation on the C++ side follows what was already existing in GATE 9, however substituting the external CLHEP library to sample from a correlated bivariate Gaussian distribution with a custom implementation. The implementation was verified on 5 test cases agreeing within tolerances with GATE v9.3 and GATE-RTion v1.
In a second step a beamline model class was introduced. As in the previous version, the energy dependent parameters of a specific beamline can be described with a polynomial function of n-th order.

In the third step, the TP source was implemented initiating several PB sources using the parameters calculated from the beamline model. Test cases created to verify absolute dose, optics and rotations of the TP source simulations, are in good agreement with GATE-RTion v1.

The PB source, ion source description (beam model) and TP source were successfully translated from Gate 9 to Gate 10 making use of the new python interface and new features were added and verified.

Orateur: Mlle Martina Favaretto (MedAustron GmbH)

Towards Gate 10 for ion beam therapy applications.pdf

11:30 Validation of TEPCActor for slab detectors in microdosimetry

Monte Carlo (MC) simulations of microdosimetric spectra are available for different codes, detectors and particle types/energies. While good results can be achieved if the detector geometry is modelled in detail [1, 2], a simpler approach can also result in very good agreement.
At five positions along a 238.6 MeV/u carbon ion pencil beam, spectra were collected using the Silicon MicroPlus probe. The measurements were collected along CAX as well as 10 mm off-center. The MC was performed with GATE/Geant4, using our validated beam model and a simplified detector geometry. Only the 400 sensitive volumes (diameter = 20 μm, thickness = 10 μm, spaced 50 μm apart) were simulated, none of the housing/surrounding materials.
The off-center measurements show a higher contribution of secondary particles, which we were able to reproduce in the MC (Figure 1). The agreement of the shape of the spectra is excellent for all positions along the Bragg peak. Especially in the fragmentation tail, where we were able to measure and simulate the boron edge with high accuracy.
This simplified detector geometry could be implemented in future GATE installation checks, to perform a validation down to the microdosimetric scale. MC also offers a spectrum down to very low keV/μm values, which can be used to complete any experimentally obtained spectrum.

[1] David Bolst et al 2020 Phys. Med. Biol. 65 045014
[2] Anna Bianchi et al 2023 Phys. Med. Biol. 68 034001

Orateur: Mlle Sandra Barna (Medizinische Universität Wien)

presentation_Barna_et_al_clean_animation.pptx

11:50 Modelling and corrections of random & scatter coincidences in J-PET

Orateur: Szymon Parzych (Jagiellonian University)

GATE_Meetinig_2023_SParzych.pptx

12:10 → 14:00 J-PET visit and lunch